error detection and correction codes are provided. For a word of m bits that is to be coded, a vector with m components, each component corresponding to a bit of the word, is formed. The vector is multiplied, using a computing circuit, by a parity control matrix. The parity control matrix includes at least one couple of complementary lines.
|
6. A system for determining r error detection bits of a word of m bits that is to be coded, comprising:
means for forming a vector with m components, each component corresponding to a bit of the word that is to be coded; and
a circuit to multiply said vector by a parity control matrix and provide said r error detection bits as a result of said multiplication, the parity control matrix including at least one couple of complementary lines.
7. A coding circuit enabling implementation of an error correction and/or detection code comprising a matrix that includes at least one couple of complementary lines, the circuit including elementary adders, a distribution of the elementary adders corresponding to the distribution of the “1s” in the matrix including lines and columns formed of elements equal to “1” or “0”, in which at least two of said lines are formed of complementary elements.
1. A method for determining r error detection bits of a word of m bits that is to be coded, including a step of forming a vector with m components, each component corresponding to a bit of the word that is to be coded, and a step of multiplying, by a circuit, said vector by a parity control matrix, the result of said multiplying step being a vector formed of said r error detection bits, wherein the parity control matrix includes at least one couple of complementary lines.
11. A method for determining r error detection bits of a word of m bits that is to be coded, including a step of forming a vector with m components, each component corresponding to a bit of the word that is to be coded, and a step of multiplying, by a circuit, said vector by a parity control matrix, the result of said multiplying step being a vector formed of said r error detection bits, wherein the parity control matrix includes at least one couple of complementary lines, wherein a binary representation of the r-2 first bits of each column indicates the column rank, except for a first column, of rank 0.
9. A decoding circuit enabling implementation of an error correction and/or detection code comprising a matrix that includes at least one couple of complementary lines, the circuit including elementary adders, a distribution of the elementary adders corresponding to the distribution of the “1s” in the matrix including
a) a first left-hand block corresponding to the parity control matrix used in the coding and;
b) a second square block of dimension r-1, in the form of a diagonal matrix including only “1s” on its diagonal,
c) a last column only including “0s” on the r-1 first lines and a “1” on the last line, and
d) under the second block, elements which are complements to two of those of the last line of the second block.
4. A method for determining r error detection bits of a word of m bits that is to be coded, including a step of forming a vector with m components, each component corresponding to a bit of the word that is to be coded, and a step of multiplying, by a circuit, said vector by a parity control matrix, the result of said multiplying step being a vector formed of said r error detection bits, wherein the parity control matrix includes at least one couple of complementary lines, wherein m is an even number, and the parity control matrix is such that:
a) a first half of a penultimate line, corresponding to the m/2 first columns, includes a “0” at each end and “1s” everywhere else,
b) a second half of the penultimate line, corresponding to the last m/2 columns, includes a “1” it at each end and “0s” everywhere else, and
c) a last line is complementary to the penultimate line.
5. A method for determining a syndrome, said method comprising a step of forming a vector with m+r components, each component corresponding to a bit of an m+r-bit word that is to be decoded, the m+r bits of said word that is to be decoded corresponding, before processing, to m bits of a word that is to be coded and r error detection bits obtained by a step of forming a vector with m components, each component corresponding to a bit of the word that is to be coded, and a step of multiplying, by a circuit, said vector by a parity control matrix, the result of said multiplying step being a vector formed of said r error detection bits, wherein the parity control matrix includes at least one couple of complementary lines, applied to the word of m bits that is to be coded, and a step of multiplying, by means of a computing circuit, said vector by a specific matrix, wherein the specific matrix includes:
a) a first left-hand block corresponding to the parity control matrix used in the coding and;
b) a second square block of dimension r-1, in the form of a diagonal matrix including only “1” on its diagonal,
c) a last column only including “0s” on the r-1 first lines and a “1” on the last line, and
d) under the second block, elements which are complements to two of those of the last line of the second block.
3. The method of
|
1. Field of the Invention
The present invention relates to error detection and/or correction codes.
Error detection codes have a technical effect and solve a significant technical problem. Indeed, they enable restoring the value of one or several erroneous bits, for example, after a storage or a transmission. Without such codes, any storage or transmission of digital data would be difficult.
The present invention relates in particular to linear codes of Hamming type.
2. Discussion of the Related Art
The Hamming code is an error detection and correction code used in many fields. A first example of a Hamming code and its use for the data storage in a memory will be described, in the case where the data to be stored are in the form of 16-bit words.
Let X be the word to be stored. X can be represented by a vector Xe, the 16 components X0 to X15 of which correspond to the 16 bits of the word to be stored. Five error detection bits C1(C0 . . . C4) are obtained by multiplying a parity control matrix H, called a Hamming matrix, of dimensions 5×16, by vector Xe in the form of a column vector.
Xj being the j-th component of vector Xe.
In write mode, 21-bit words, formed of the 16 data bits Xj and of the 5 detection bits C1, are written into the memory. In read mode, the read word includes 16 bits Xr corresponding to the data bits and 5 bits Cr corresponding to the detection bits. It is possible for Xr and Cr not to be equal to Xj and Ci if errors have occurred between the write and read operations.
To detect and/or correct possible errors on the read bits, a syndrome S with five components S0, . . . S4 is calculated by multiplying a determined matrix H′ of dimensions 5×21 by a column vector with 21 components, including the 16 bits Xr and the 5 bits Cr.
If syndrome S has all its components equal to 0, the storage occurs with no error and all the bits of the read word, be they data bits or detection bits, are correct.
If S is different from 0, the read word includes one or several errors. If a single bit of the read word is erroneous, the obtained syndrome enables correcting the error. Indeed, the syndrome corresponds in this case to the column having had its elements multiplied by the erroneous bit. Thus, if the calculated syndrome is equal to:
the components (00011) of the syndrome correspond to the elements of the first column of the Hamming matrix, which means that the first bit, X0, is erroneous.
Similarly, if the calculated syndrome is equal to:
and there is a single error in the read word, this means that the first detection bit C0 is erroneous.
The above-described Hamming code cannot detect two errors. Thus, if an error has occurred on bits X1 and X2, the obtained syndrome is equal to the sum modulo 2 of the syndromes corresponding to errors on X1 and X2, that is, to: S′″=(00101)+(00110)=(00011). The obtained syndrome indicates an error on bit X0, which is wrong.
Indeed, the above Hamming code is known to have a minimum code distance d=3 and a linear code like the Hamming code is known to be able to correct L errors and to detect L+1 errors if its minimum code distance d is strictly greater than 2L+1.
To improve the above code and turn it into a code having a minimum code distance equal to 4, it is known to add to the word to be stored a total parity bit P.
Total parity bit P is calculated by adding modulo 2 all the data bits and all the detection bits. The total parity bit is added to the word to be stored, and the word to be stored, the detection bits, and the total parity bit are altogether stored.
In read mode, the read word is multiplied by parity control matrix H″ shown in
The obtained syndrome S′ is illustrated in
The code thus obtained is a so-called “SEC-DED” code (“Single Error Correction”—“Double Error Detection”). This code has a minimum code distance equal to four and can detect two errors in all cases, two errors being indicated by the fact that the last component of the syndrome, S5, is zero while the syndrome is different from the zero vector. However, the above code has several disadvantages.
Thus, upon coding, the calculation of total parity bit P is required. This calculation requires a large number of adders, since all data bits and detection bits must be added modulo 2. Further, the calculation of the total parity bit cannot be performed in parallel with the calculation of the detection bits, since it requires the previous knowledge of the detection bits. Accordingly, it must be awaited that all detection bits have been calculated to calculate total parity bit P, which wastes time.
Upon decoding, the calculation of the last syndrome component, S5, requires a high number of adders, and the decoding circuit has a low compactness. Further, since each addition lasts for some time, the calculation of the last syndrome component has a relatively long duration and the decoding is not optimal.
It should also be noted that, in the above-described Hamming code, the Hamming matrix is neither symmetrical, nor regular. Thus, considering that the elements of a column correspond to the binary representation of a number, the variation of this number is not regular and includes jumps. This makes difficult the forming of a circuit implementing the Hamming code as well as the syndrome decoding.
An object of the present invention is to provide an error detection and/or correction code which is simpler than the corresponding Hamming code.
Another object of the present invention is to provide a method and a device for implementing such an error detection and/or correction code.
Another object of the present invention is to provide a method and a device for implementing an error detection and/or correction code that can be implemented in a simple way by an integrated circuit.
To achieve these objects as well as others, the present invention provides a method for determining r error detection bits of a word of m bits to be coded, including the step of calculating the product of a vector with m components representative of said word of m bits to be coded and of a parity control matrix. The parity control matrix includes at least one couple of complementary lines.
According to an embodiment of the present invention, the binary representation of the r−2 first bits of each column indicates the column rank, except for the first column, of rank 0.
According to an embodiment of the present invention, in is an even number, and the parity control matrix is such that:
a) the first half of the penultimate line, corresponding to the m/2 first columns, includes a “0” at each end and “1s” everywhere else,
b) the second half of the penultimate line, corresponding to the last m/2 columns, includes a “1” at each end and “0s” everywhere else, and
c) the last line is complementary to the penultimate line.
According to an embodiment of the present invention, two or more than two lines and/or columns are permuted.
According to an embodiment of the present invention, number r of the detection bits is equal to n+2, n being the number of bits necessary to binarily represent number m of bits of the word to be coded.
The present invention also relates to a method for determining a syndrome representative of possible errors having occurred in a processing to an m+r-bit word, the m+r bits corresponding, before processing, to m bits of a word to be coded and r error detection bits obtained by a method such as described hereabove applied to the word of m bits to be coded, the syndrome being obtained in a step consisting of multiplying a specific matrix by a vector having m+r components representative of the m+r-bit word. The specific matrix includes:
a) a first left-hand block corresponding to the parity control matrix used in the coding and;
b) a second square block of dimension r−1, in the form of a diagonal matrix including only “1s” on its diagonal,
c) a last column only including “0s” on the r−1 first lines and a “1” on the last line, and
d) under the second block, elements which are complements to two of those of the last line of the second block.
The present invention also relates to an error correction and/or detection code using a matrix such as defined hereabove.
The present invention also relates to a coding circuit enabling implementation of an error correction and/or detection code such as defined hereabove, including elementary adders, the distribution of the elementary adders corresponding to the distribution of the “1s” in a matrix such as defined hereabove in the method for determining r error detection bits.
The present invention also relates to a decoding circuit enabling implementation of an error correction and/or detection code such as defined hereabove, including elementary adders, the distribution of the elementary adders corresponding to the distribution of the “1s” in a matrix such as defined hereabove in the method for determining a syndrome.
According to an embodiment of the present invention, the elementary adders are formed by gates of XOR type.
The foregoing objects, features and advantages of the present invention will be discussed in detail in the following non-limiting description of specific embodiments in connection with the accompanying drawings.
Matrix M is formed as follows. The first column of the matrix corresponds to rank 0. For the other columns, elements are placed on the first four lines, the binary representation of which indicates the column rank. Thus, the second column, of rank 1, includes on the first four lines elements “0001”, which are the binary representation of number 1. The sixteenth column (rank 15) includes on the first four lines elements 1111, which are the binary representation of number 15. The first four elements of the first columns have been chosen to be equal to “0011”, which correspond to the first four elements of the column of rank 3. This choice is not critical and, for the first four elements of the first column, the first four elements of any column of matrix M could have been taken, as will be seen hereafter, provided that they include more than a single “1”.
After the first four lines, the fifth and penultimate lines of matrix M are formed. For this purpose, matrix M is divided in two by a median axis xx′ separating the 16 columns in 8 left-hand columns (rank 0 to 7) and 8 right-hand columns (rank 8 to 15). The first half of the fifth line, corresponding to a block K1 of 8 elements to the left of axis xx′, is formed by placing a “0” at each end of block K1 and “1s” everywhere else. Thus, the first half of the fifth line has expression “01111110”. To form the second half of the fifth line, corresponding to a block K2, a “1” is placed at each end of block K2 and “0s” are placed everywhere else. The fifth line thus has as full expression “0111111010000001”.
The elements of the last and sixth line L of matrix M are the complements to two of the elements of the fifth line. Line L thus has as expression “1000000101111110”.
The matrix M thus formed has distinct columns linearly independent two by two. When matrix M is multiplied by a column vector of sixteen components corresponding to the bits of the word to be coded, six detection bits C0–C5 are obtained, which are added to the word to be coded to form a 22-bit coded word. The decoding of the coded word enables correction of an error and detection of two errors.
It should be noted that matrix M includes a line couple, the last and the penultimate lines, which are complementary to each other. This provides a significant advantage to the code of the present invention. In particular, the provided code thereby avoids calculation of a total parity bit upon coding. The coding circuit is thus more compact and includes less adders than in prior art. All the bits added to the word to be coded can be calculated in parallel, which results in a great time gain.
It should be noted that, like for matrix M, the last two lines of matrix M′ are complementary. Thus, if the sum modulo 2 of the last two syndrome components, S4 and S5, is calculated, the sum modulo 2 of all data and detection bits of the word to be decoded, that is, a total parity bit Pr, is obtained by a simple calculation substantially using twice as few adders and as less time as in the case of the corresponding Hamming code. Upon decoding, total parity bit Pr, obtained without calculation of a total parity bit upon coding, is used as follows.
If the obtained syndrome is equal to the zero vector, there is no error, either on the data bits, or on the detection bits.
If the obtained syndrome is different from the zero vector and total parity bit Pr is equal to 1, this means that there is a single error and it can be corrected. Indeed, the syndrome components, in this case, correspond to the elements of the column of matrix M′ corresponding to the erroneous bit. Further, since the first four syndrome components indicate the rank of the erroneous bit except for the first one, it is very easy to determine the position of the erroneous bit and to correct it. This is an additional advantage over the Hamming code.
If the syndrome is different from the zero vector and total parity bit Pr is equal to 0, two errors are present, which are detected.
As shown in
When there is no adder at the intersection of column i and of line j, this means that column i and line j cross with no influence upon each other. This means that the bit provided to the involved input is not used to calculate component Sj of the syndrome.
An additional column a, located to the left of column 0, connects input ej of each first adder of a line to a grounded line (GND).
The operation of the decoding circuit will be explained for the calculation of component S3 of the syndrome, corresponding to the line of rank 3. Starting from the left, the first encountered adder is adder G0,3. Input e3 of adder G0,3 is connected to ground and its input e0 receives data bit X′0 via input E0 of the circuit. At the output of adder G0,3, s=0⊕X′0, that is, X′0. The signal provided by adder G0,3 drives input e3 of adder G1,3, which calculates X′0⊕X′1. The calculation carries on in this way until adder G19,3 is reached, which adds modulo 2 the result provided by adder G15,3 and detection bit C′3. Thus:
S3=X′0⊕X′1⊕X′3⊕X′5⊕X′7⊕X′9⊕X′11⊕X′15 ⊕C′3,
which effectively corresponds to the multiplication of the fourth line of matrix M′ by a vector having as components the bits of the word to be decoded. Generally, the decoding circuit of
In the code described in relation with
The circuit used for the coding corresponds to the decoding circuit, except for the last six columns, which do not exist for the coding. The outputs of the coding circuit provide the detection bits.
Of course, it is easy to generalize matrix M of
To generalize matrix M, number r of necessary detection bits is first determined. Then, matrix Mm used for the coding is built, so that the first r−2 elements of each column of even rank indicate, except for the first column, the column rank (it should be reminded that the rank of the first column is 0, and that of the last columns is m−1). To form the penultimate line, the line is separated in two portions of m/2 elements each. The first portion of the penultimate line (columns 0 to [m/2−1]) includes a “0” at each end and “1s” everywhere else. The second portion of the penultimate line (columns m/2 to m−1) includes a “1” at each end and “0s” everywhere else.
The last line of matrix Mm is complementary to the penultimate line, that is, the elements of the last line are the complements to 2 of the elements of the penultimate line.
The first r−2 elements of the first column of matrix Mm may be identical to the r−2 elements of any column, provided that the following conditions are respected: the first column must be different from all other columns; its first four elements must include at least two “1s”, to be different from the columns of the matrix used upon decoding.
To form matrix M′m used for the decoding, parity control matrix Mm is first taken, and completed to the right with a square sub-matrix R of dimension r×r. Sub-matrix R includes “1s” on its main diagonal, and “0s” everywhere else except on its last line, the elements of which are the complements of those of the penultimate line of sub-matrix R. The last line of sub-matrix R thus includes “1s” everywhere except at the penultimate column.
The code using matrixes Mm and M′m has a minimum code distance equal to four. It enables correcting one error and detecting two errors. Upon decoding, the obtained syndrome has r components. A total parity bit is obtained by adding modulo 2 the last two syndrome components.
If the syndrome is the zero vector, there is no error. If the syndrome is different from the zero vector and the total parity bit is equal to “1”, there is a single error. This error is easily corrected, since the first r−2 components of the syndrome indicate the rank of the erroneous bit, except for rank 0. If the syndrome is different from the zero vector and the total parity bit is equal to “0”, two errors are present.
It should be noted that, like for matrix M′, the last two lines of matrix M′m are complementary. The sum modulo 2 of the last two components of the obtained syndrome represent the sum modulo 2 of each of the data bits and of the detection bits, that is, a total parity bit, obtained without having had to calculate a total parity bit upon coding.
In
In
Of course, the present invention is likely to have various alterations, modifications, and improvements which will readily occur to those skilled in the art. In particular, the matrixes according to the present invention described hereabove are examples only and those skilled in the art may easily modify them. Thus, any line and/or column permutation in a matrix according to the present invention described hereabove is within the scope of the present invention.
The two complementary lines of the used matrixes are not necessarily formed by the elements described hereabove. They are not necessarily consecutive either.
Also, the matrixes used for the coding and/or the decoding may include more than one couple of complementary lines, if desired.
Number N of bits of the word to be coded may be even or odd, the matrixes used upon coding and/or decoding including at least one couple of two complementary lines. If number N is odd, a matrix Mm such as described hereabove with an even m equal to N+1 may for example first be formed. Then, matrix MN to be used upon coding may easily derive from matrix Mm by eliminating any column, for example, the first one or the last one.
Also, the present invention has mainly been described in the context of the storage of words in a memory. Of course, the present invention also applies to any coding and decoding of words to which an error detection and/or correction code is desired to be assigned. For example, the present invention applies to transmission.
Such alterations, modifications, and improvements are intended to be part of this disclosure, and are intended to be within the spirit and the scope of the present invention. Accordingly, the foregoing description is by way of example only and is not intended to be limiting. The present invention is limited only as defined in the following claims and the equivalents thereto.
Patent | Priority | Assignee | Title |
7849388, | Sep 22 2005 | ROHM CO , LTD | Signal decoding method and device, and signal storage system |
Patent | Priority | Assignee | Title |
5751744, | Feb 01 1993 | RPX Corporation | Error detection and correction circuit |
5802078, | Nov 27 1992 | Kokusai Electric Co., Ltd. | Error detector for error detecting codes |
6336203, | May 30 1995 | Mitsubishi Denki Kabushiki Kaisha | Error correction coding and decoding method, and circuit using said method |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Oct 03 2002 | STMicroelectronics S.A. | (assignment on the face of the patent) | / | |||
Oct 21 2002 | MURILLO, LAURENT | STMICROELECTRONICS S A | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 013511 | /0095 |
Date | Maintenance Fee Events |
Oct 27 2010 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
Oct 23 2014 | M1552: Payment of Maintenance Fee, 8th Year, Large Entity. |
Oct 25 2018 | M1553: Payment of Maintenance Fee, 12th Year, Large Entity. |
Date | Maintenance Schedule |
May 29 2010 | 4 years fee payment window open |
Nov 29 2010 | 6 months grace period start (w surcharge) |
May 29 2011 | patent expiry (for year 4) |
May 29 2013 | 2 years to revive unintentionally abandoned end. (for year 4) |
May 29 2014 | 8 years fee payment window open |
Nov 29 2014 | 6 months grace period start (w surcharge) |
May 29 2015 | patent expiry (for year 8) |
May 29 2017 | 2 years to revive unintentionally abandoned end. (for year 8) |
May 29 2018 | 12 years fee payment window open |
Nov 29 2018 | 6 months grace period start (w surcharge) |
May 29 2019 | patent expiry (for year 12) |
May 29 2021 | 2 years to revive unintentionally abandoned end. (for year 12) |